Abacn opened a new issue, #26408:
URL: https://github.com/apache/beam/issues/26408
### What happened?
When implementing BigQuery STORAGE_WRITE_API performance test, it is found
that withAvroFormatFunction needs to be set differently:
For FILE_LOAD:
```
BigQueryIO.<byte[]>write()
.withAvroFormatFunction(
writeRequest -> {
byte[] data = writeRequest.getElement();
GenericRecord record = new
GenericData.Record(writeRequest.getSchema());
record.put("data", ByteBuffer.wrap(data));
return record;
})
```
See
https://github.com/apache/beam/blob/44a17cbe429699160e689bd88935c32c6c2de6b2/sdks/java/io/bigquery-io-perf-tests/src/test/java/org/apache/beam/sdk/bigqueryioperftests/BigQueryIOIT.java#L169
For STORAGE_WRITE_API, however, it needs to be
```
BigQueryIO.<byte[]>write()
.withAvroFormatFunction(
writeRequest -> {
byte[] data = writeRequest.getElement();
GenericRecord record = new
GenericData.Record(writeRequest.getSchema());
record.put("data", data);
return record;
})
```
Using `ByteBuffer.wrap(data)` will throw this Exception at run time:
```
Caused by: java.lang.ClassCastException: class java.nio.HeapByteBuffer
cannot be cast to class [B (java.nio.HeapByteBuffer and [B are in module
java.base of loader 'bootstrap')
at
org.apache.beam.sdk.io.gcp.bigquery.AvroGenericRecordToStorageApiProto.lambda$static$4(AvroGenericRecordToStorageApiProto.java:90)
at
org.apache.beam.sdk.io.gcp.bigquery.AvroGenericRecordToStorageApiProto.scalarToProtoValue(AvroGenericRecordToStorageApiProto.java:375)
at
org.apache.beam.sdk.io.gcp.bigquery.AvroGenericRecordToStorageApiProto.toProtoValue(AvroGenericRecordToStorageApiProto.java:332)
at
org.apache.beam.sdk.io.gcp.bigquery.AvroGenericRecordToStorageApiProto.toProtoValue(AvroGenericRecordToStorageApiProto.java:318)
at
org.apache.beam.sdk.io.gcp.bigquery.AvroGenericRecordToStorageApiProto.messageValueFromGenericRecordValue(AvroGenericRecordToStorageApiProto.java:296)
at
org.apache.beam.sdk.io.gcp.bigquery.AvroGenericRecordToStorageApiProto.messageFromGenericRecord(AvroGenericRecordToStorageApiProto.java:183)
at
org.apache.beam.sdk.io.gcp.bigquery.StorageApiDynamicDestinationsGenericRecord$GenericRecordConverter.toMessage(StorageApiDynamicDestinationsGenericRecord.java:73)
at
org.apache.beam.sdk.io.gcp.bigquery.StorageApiConvertMessages$ConvertMessagesDoFn.processElement(StorageApiConvertMessages.java:139)
```
at here:
https://github.com/apache/beam/blob/44a17cbe429699160e689bd88935c32c6c2de6b2/sdks/java/io/google-cloud-platform/src/main/java/org/apache/beam/sdk/io/gcp/bigquery/AvroGenericRecordToStorageApiProto.java#L90
In contrast, if directly feed a byte[] for FILE_LOAD, it gives the following
error:
```
Caused by: org.apache.avro.file.DataFileWriter$AppendWriteException:
org.apache.avro.AvroRuntimeException: Unknown datum type [B: [B@6f10f22c
at
org.apache.avro.file.DataFileWriter.append(DataFileWriter.java:317)
at
org.apache.beam.sdk.io.gcp.bigquery.AvroRowWriter.write(AvroRowWriter.java:52)
at
org.apache.beam.sdk.io.gcp.bigquery.WriteBundlesToFiles.processElement(WriteBundlesToFiles.java:236)
Caused by: org.apache.avro.AvroRuntimeException: Unknown datum type [B:
[B@6f10f22c
at
org.apache.avro.generic.GenericData.getSchemaName(GenericData.java:933)
at
org.apache.avro.generic.GenericData.resolveUnion(GenericData.java:892)
at
org.apache.avro.generic.GenericDatumWriter.resolveUnion(GenericDatumWriter.java:307)
```
Ideally we should support ByteBuffer in AvroGenericRecordToStorageApiProto.
### Issue Priority
Priority: 2 (default / most bugs should be filed as P2)
### Issue Components
- [ ] Component: Python SDK
- [X] Component: Java SDK
- [ ] Component: Go SDK
- [ ] Component: Typescript SDK
- [X] Component: IO connector
- [ ] Component: Beam examples
- [ ] Component: Beam playground
- [ ] Component: Beam katas
- [ ] Component: Website
- [ ] Component: Spark Runner
- [ ] Component: Flink Runner
- [ ] Component: Samza Runner
- [ ] Component: Twister2 Runner
- [ ] Component: Hazelcast Jet Runner
- [ ] Component: Google Cloud Dataflow Runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
